Chapter 19: Obstruction

In part 2, I explained how purposefully difficult-to-use UI design (sludge) can be employed as part of an exploitative resource depletion strategy, to make users fatigued and give up trying to achieve their goals, or to soften them up prior to a bigger deception. This is exactly what the obstruction category of deceptive pattern is all about.

Obstruction in Facebook and Google’s privacy settings

When the GDPR came into effect in the EU, companies were obliged to change the way they presented their privacy options to ensure that users had the means to consent (or dissent) to proposed uses of their personal data. This consent needed to be ‘freely given, specific, informed and unambiguous’ (article 4(11)).1

On behalf of its citizens, in 2018 the publicly funded Norwegian Consumer Council (Forbrukerrådet) carried out an investigation into this practice.2 It found that Facebook and Google had used deceptive patterns in their user interfaces to ‘nudge users away from the privacy-friendly choices’. They did this by using obstruction: they made it easy to accept the privacy-invading settings and hard to reject them. You can see this in the figure below. It’s one click to ‘accept and continue’, but if the user wants to ‘reject and continue’, there is no equivalent button. Instead, they have to click the ambiguously labelled ‘manage data settings’ button, and then they have to push an ambiguous toggle to the left. Notice that the toggle is improperly labelled – the user is not clearly told whether they have successfully rejected the ad tracking.3

Screenshot of three mobile screens showing Facebook’s personal data settings in 2018. After the user clicks the prominent blue ‘Get started’ button on the first data settings screen, the next screen shows a summary of and links to Facebook’s data privacy policy; at the bottom, the ‘Accept and continue’ button shares the prominence of the earlier ‘Get started’ button, but the button to ‘Manage data settings’ is ambiguously labelled and shaded white, the same as the background colour. The final screen, reached by pressing the ‘Manage data settings’ button, shows a toggle control under the heading ‘Ads based on data from partners’ with the label ‘Allowed: Ads will be more relevant to you’. The toggle’s default setting is slid to the right, but there is no indication that sliding it to the left will prevent a user’s personal data being shared.
Facebook’s data settings (Forbrukerrådet, 2018).

Google’s approach was similar. Google required users to sign in first, and to then look for and use the privacy settings dashboard of their own volition. From there, the user could choose to opt out. Again, this is obstruction, and it is the opposite of what the GDPR regulation requires.4

Screenshot of Google’s ad personalisation settings. To the right of a heading ‘Ads Personalisation on Google Search’ and a label ‘See more useful ads when you’re using Google Search’ is a toggle control, positioned to the right; no indication of what happens if slid to the left is given. At the bottom of the screen under the heading ‘Ads Personalisation Across the Web’ and the label ‘See more useful ads on YouTube and 2+ million websites that partner with Google to show ads’ are two orange buttons labelled ‘TURN ON’ and ‘TURN OFF’  in the same font at the same weight and colour.
Google’s data settings (Forbrukerrådet, 2018).

In both examples, users were required to expend more effort and attention to opt out than to just go along with the defaults and be automatically opted in. The Norwegian Consumer Council argued this was an example of the default effect bias being employed as a deceptive pattern for commercial gain.

In 2018, the Norwegian Consumer Council filed a complaint against Google on this matter. Five years later, they are still waiting for a final decision by the Irish Data Protection Commissioner.

In 2022, ten European consumer groups filed a second complaint against Google for employing similar tactics but with a greater focus on location data and the Google account sign up process.5 As BEUC writes: ‘Tech giant Google is unfairly steering consumers towards its surveillance system when they sign up to a Google account, instead of giving them privacy by design and by default as required by the General Data Protection Regulation (GDPR).’

The hard to cancel deceptive pattern

Hard to cancel is a type of obstruction that involves businesses making it difficult for users to cancel their subscriptions. It is often paired up with a very easy and frictionless subscription experience – making it easy to join and hard to leave. When seen in this pairing, it is sometimes referred to as a ‘roach motel’ – a humorous reference to a pest control device of the same name.6

Hard to cancel by the New York Times

The New York Times has published a number of articles on deceptive patterns over the years, and it holds a progressive view on consumer rights and regulation. However, until recently, it did not extend that view to its own digital services, and it is famous for using the hard to cancel deceptive pattern.

As Twitter user @vanillatary succinctly put it: ‘this should literally be illegal […] the extra subscriptions they keep by making it annoying and time consuming to unsubscribe are clearly worth far more than the cost of hiring additional staff to handle the unsubscribe calls, which could’ve been handled 100 times more efficiently by a few lines of web code […] So some % of the NYT’s business model is based on holding on to paid customers who no longer actually consider their product worth the cost’.7

The screenshots below show an easy-to-follow subscription process on the one hand, and an obstructive cancellation process on the other. Instead of giving the user a button to directly cancel a subscription, the NYT provided instructions on contacting customer services.8

Screenshot of the New York Times subscription form. Under the heading ‘SPECIAL OFFER Unlimited access to all the journalism we offer’ and the price is a clear black button with ‘SUBSCRIBE NOW’ in white text; below the button is the message ‘Cancel or pause anytime’.
A typical New York Times subscription upsell page in November 2021, provided by Twitter user @vanillatary.
Screenshot of the New York Times ‘Cancel your subscription’ page. The text states ‘There are several ways to unsubscribe from The Times.’ Further information given is ‘Speak with a Customer Care Advocate’ with a telephone number and working hours provided; alternatively a user can ‘Chat with a Customer Care Advocate’ using an online chat service. These options to cancel are much more difficult for the user than how to subscribe.
The New York Times subscription cancellation page, provided by Twitter user @vanillatary in November 2021.

You may be wondering if 7 minutes is really that long. Well, if we compare it to the approximately 500 milliseconds it would have taken to submit the same request via a button on a web page, it’s over 800 times longer. That’s a big difference.

In September 2021, the New York Times fell foul of a class action lawsuit on issues relating to its automatic renewal experience. Mainly, its transgressions fell under California’s Automatic Renewal Law, which states among other things:10

‘a business that allows a consumer to accept an automatic renewal or continuous service offer online shall allow a consumer to terminate the automatic renewal or continuous service exclusively online, at will, and without engaging any further steps that obstruct or delay the consumer’s ability to terminate the automatic renewal or continuous service immediately.’

This suit mandated the payment of over $5 million in settlement fees.11 For a period, the newspaper added the means to cancel online for California residents, then in early 2023 it extended this for all customers, regardless of their location. In a similar class action lawsuit, the Washington Post was found to be deploying similar deceptive patterns, making it difficult for users to cancel subscriptions.12 It too had to pay a settlement, this time for $6.8 million. The FTC also brought a similar complaint against online education provider ABCmouse, in which ‘ABCmouse didn’t clearly tell parents that their subscriptions would renew automatically, and then the company made it very difficult for them to cancel’.13 The company had to pay $10 million in damages, and – thankfully – was also told to stop employing these deceptive patterns.

In February 2021, someone published a transcript of their NYT live chat cancellation experience, which showed the process had taken 17 minutes, start to finish.9 I tried it too, here in the UK, on 13 November 2021. It took me 7 minutes to cancel my subscription. The chat representative asked me why I was cancelling, and then used my reason as a hook to start what I presume was a scripted attempt to change my mind.

In 2021, sixteen consumer organisations in the EU and the US filed complaints against Amazon for employing the hard to cancel deceptive pattern against users who wished to cancel their Amazon Prime subscriptions. In a report titled ‘You can log out but you can never leave’, the Norwegian Consumer Council explained the process of cancellation, which involved a labyrinth of confusing choices, shown below.14

A collection of screenshots showing the complex process of cancelling an Amazon Prime subscription. 1. Opening the navigation menu. 2. Tapping ‘Account’. 3. Choosing ‘Prime membership’. 4. Tapping ‘Manage membership’ above a screen prompting the benefits of membership’. 5. Pushing a button marked ‘End membership’. 6. The next screen emphasises what the user will lose by unsubscribing.
Steps 1-6 in cancelling an Amazon Prime subscription (Norwegian Consumer Council, 2021).
A collection of screenshots showing the complex process of cancelling an Amazon Prime subscription. 7. A button labelled ‘Cancel My Benefits’ is shown beneath a ‘Keep My Benefits’ button. 8. The next screen again mentions what will be lost by the user and presents an offer of reduced cost if the user switches to paying annually. 9 The button to ‘Continue to Cancel’ is shown below a button to ‘Keep My Membership’. 10. Under the heading ‘We’re sorry to see you go. Please confirm the cancellation of your membership’, the user is given the option to ‘pause’ their subscription instead. 11. The user has to scroll farther down to finally reach the ‘End Now’ button at the bottom of the screen. 12. The user reaches the cancellation confirmation screen, which also states that ‘You may rejoin Prime at any time’ and shows what’s included with Prime membership.
Steps 7-12 in cancelling an Amazon Prime subscription (Norwegian Consumer Council, 2021).

In response to the slew of complaints, the European Commission ran an investigation that they concluded by instructing Amazon to change the cancellation procedure for all European consumers.15 The European Commissioner gave a strongly worded statement, saying:

‘Opting for an online subscription can be very handy for consumers as it is often a very straightforward process, but the reverse action of unsubscribing should be just as easy. Consumers must be able to exercise their rights without any pressure from platforms. One thing is clear: manipulative design or “dark patterns” must be banned.’

At roughly the same time, seven US consumer groups wrote a letter to the FTC, asking them to investigate – and it did.16 During the investigation, internal documents were leaked to the business news site Insider.com, who revealed that Amazon had knowingly implemented deceptive patterns as a part of a strategy to retain customers and reduce churn.17 Leaked documents admitted things like: ‘The button’s label “Continue with FREE 1-day shipping” did not adequately convey that the customer was signing up for a membership’; ‘Unintentional sign-ups erode customer trust’; and ‘Either way, improvements in clarity during sign-up are needed’. Furthermore, data leaked in August 2017 also showed that 67% of the cancellation requests directly handled by the Prime team were related to ‘accidental sign-ups’. In other words, the evidence implies that Amazon knew exactly what it was doing and continued anyway.

The FTC’s Amazon investigation still seems to be ongoing at the time of writing. In March 2023, the FTC proposed some new rules on the hard to cancel deceptive pattern. ‘The proposed rule would require that companies make it as easy to cancel a subscription as it is to sign up for one’ said FTC Chair Lina M Khan.18 This was even supported by President Biden, who tweeted: ‘I support @FTC’s proposal requiring companies to fix this. It shouldn’t be harder to cancel a service than it was to subscribe for it’.19

Amazingly, even after all this, cancelling an Amazon Prime subscription is still not as easy as it could be. The fact that Amazon is willing to spend so much money in legal proceedings and take on this much risk gives us a clue as to how profitable this practice must be.

Buy the book...

Since 2010, Harry Brignull has dedicated his career to understanding and exposing the techniques that are employed to exploit users online, known as “deceptive patterns” or “dark patterns”. He is credited with coining a number of the terms that are now popularly used in this research area, and is the founder of the website deceptive.design. He has worked as an expert witness on a number of cases, including Nichols v. Noom Inc. ($56 million settlement), and FTC v. Publishers Clearing House LLC ($18.5 million settlement). Harry is also an accomplished user experience practitioner, having worked for organisations that include Smart Pension, Spotify, Pearson, HMRC, and the Telegraph newspaper.